Thermal Control of Brown Adipose Tissue in ¹⁸F-FDG PET

• hypermetabolic brown adipose tissue
• BAT, ¹⁸F-FDG PET
• temperature control

Brown adipose tissue (BAT) has been shown to exist in both pediatric and adult humans (1) and is responsible for nonshivering thermogenesis. In response to cold conditions, nonshivering thermogenesis may be initiated, and heat is produced within the fat cells to maintain core body temperature (2). BAT is characterized by a higher degree of vascularity and mitochondrial density over white adipose tissue, and it is these characteristics that define both coloration and function (3). Metabolically, BAT is designed simply to generate heat through a process initiated by the sympathetic nervous system (3). These characteristics explain the use of both metaiodobenzylguanidine and sestamibi (and tetrofosmin) for imaging BAT (3). BAT uses glucose as a source of adenosine triphosphate to generate the heat, providing a pathway by which ¹⁸F-FDG can localize within the brown adipocyte (4).

BAT is reported to be present on approximately 2%–4% of ¹⁸F-FDG PET scans (2,5,6) predominantly in 5 locations: the cervical, supraclavicular, paraaortic, paravertebral, and suprarenal areas (7). During the winter, it is more commonly noted in women and in the pediatric population, where it plays a major role in thermal control (2). Nonetheless, BAT is not limited to these cohorts of patients and has been shown to exist in both sexes, across varying age groups, climates, and patient body mass indexes (BMIs) (8).

The emergence of PET/CT prompted more careful evaluation of lesions in the neck and chest, resulting in a recognition that nonmalignant accumulation of ¹⁸F-FDG occurred in areas of fat. When present, BAT can reduce the efficacy of ¹⁸F-FDG PET in a variety of cancers because of the difficulties, even with CT coregistration, in differentiating disease from BAT accumulation. As a result, false-positive and false-negative results may arise, and this is particularly concerning in staging of nodal disease (9). Consequently, several methods have been explored to reduce or eliminate BAT accumulation on ¹⁸F-FDG PET scans.

Several pharmaceutical methods have been used to reduce the appearance of BAT on ¹⁸F-FDG PET scans. Despite reports of some success, there are clearly risks associated with prescription medications that, combined with potential side effects, contraindications, and adverse reactions, may make a universal protocol less than optimal. To date, the administration of propranolol (a nonselective β-blocker) has proven the most effective pharmaceutical, reducing the appearance in approximately 90% of patients (10–12). It works by blocking epinephrine and norepinephrine on both β-receptors and, thus, prevents the stimulation of BAT to produce heat. Diazepam has been used previously in ¹⁸F-FDG PET to reduce neck muscle accumulation associated with stress, anxiety, and muscle tension. It is both an anxiolytic and a muscle relaxant that, due to its lipid solubility, is largely stored in adipose tissue. Reserpine is an antipsychotic that blocks norepinephrine, and thus it is not surprising that its impact on BAT accumulation is similar to that of propranolol, with reports of reductions in uptake down to 30% (13).

The limitations in applying a universal pre–¹⁸F-FDG medication protocol, the relatively small proportion of patients who would benefit from preventative measures, and the prevalence of BAT accumulation in pediatric and female populations demand an alternative noninterventional approach. Rescanning the 2.5%–30% (6,14) of patients with BAT accumulation is neither time- nor cost-effective and includes additional radiation exposure. Given the predilection of BAT toward winter months (9,15) and reports of BAT visualization in patients exposed to colder temperatures over shorter periods (2 h) (8), more careful management of the patients’ environmental temperature may provide an effective, low-cost, noninvasive method of BAT control suitable universally across patient cohorts.

Several small studies have reported that warming the patient before ¹⁸F-FDG injection and during the uptake phase may reduce BAT in up to 90% of patients (16–18). Conclusions are largely constrained by small numbers of recruits, and thus, there is a pressing need for investigation of the impact of warming on BAT accumulation in a larger cohort of patients in a variety of environmental conditions (indigenous climates). This investigation examines incident cases of and characterization of BAT in a relatively temperate climate.

RESULTS

Of the 300 patients, 183 (61%, with a 95% CI of 55.4%–66.3%) were male and 117 (39%, with a 95% CI of 33.7%–44.6%) were female. There was a statistically significant increase in the number of male patients referred for oncology-based PET (P < 0.001), and this is supported by an absence in overlap between the 2 cited CIs. The mean patient age was 59.3 y (95% CI, 57.3–61.4 y), with a range of 1–89 y and a median of 63 y. Only 12 patients were less than 18 y old (2%). The mean patient height was 173.3 cm (95% CI, 171.0–173.6 cm), the mean weight was 79.3 kg (95% CI, 77.1–81.6), and the mean BMI was 26.6 (95% CI, 25.9–27.2). The mean maximum daily temperature throughout the duration of the study was 25.7°C, with a range of 18.9°C–40.3°C. The mean minimum daily temperature was 14.2°C, with a range of 3.8°C–23.4°C.

There was a degree of heterogeneity between the 2 cohorts. Group A (not warmed) consisted of 150 patients (88 male, 62 female), and group B consisted of 150 patients (95 male, 55 female), although there was no statistically significant difference in the sex distributions between cohorts (P = 0.407). Table 1 summarizes the patient demographics within the 2 groups. The differences in demographics were not deemed statistically significant (P > 0.05). The mean age for group B was 61.0 y (95% CI, 58.1–63.9), and the mean age for group A was 57.7 (95% CI, 54.8–60.6). No statistically significant difference was noted between groups for age (P = 0.117), and this is supported by the overlap of the 95% CI. Similarly, no statistically significant differences were noted between the groups for height (P = 0.381), weight (P = 0.325), or BMI (P = 0.098). The mean maximum daily temperature experienced by patients was 22.3°C for group A and 29.2°C for group B (P < 0.0001). The mean minimum daily temperature was 9.5°C for group A and 19.0°C for group B (P < 0.0001). The differences in the mean maximum and minimum daily temperatures experienced by both groups were deemed statistically significant, and there was no overlap between 95% CIs (Table 1).

View this table:

• View inline
• View popup

TABLE 1

Summary of Respective Patient Demographics for the 2 Cohorts

BAT was present in 28 patients (9.3%) (95% CI, 6.5%–13.2%), with the remaining 272 patients (90.7%) having no evidence of BAT (95% CI, 86.8%–93.5%). Of the 150 warmed patients (group B), 3.3% (5 male) were found to have BAT present, whereas 15.3% (15 male, 8 female) of the 150 patients who were not warmed (group A) were found to have BAT (P = 0.0005) (Fig. 1; Table 2).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Numbers of patients with BAT between warmed and not-warmed cohorts, highlighting impact of warming overall and on men and women, respectively.

View this table:

• View inline
• View popup

TABLE 2

Summary of BAT Statistics

Generally, a greater proportion of men (10.9%) than women (6.8%) were found to have BAT present (Table 2); however, the difference was not deemed statistically significant (P = 0.310). The mean age of patients with BAT was 45.4 y (95% CI, 38.9–51.9)—a statistically significant decrease (P < 0.001) compared with the mean age (60.8 y) of those without BAT (95% CI, 58.7–62.9). The lack of overlap between the 95% CIs supports this observation. No statistically significant differences were noted between the patients with BAT and those without BAT with respect to height (P = 0.426), weight (P = 0.580), or BMI (P = 0.330).

Of the 12 pediatric patients (<18 y of age) included in this study, 3 (25%) were identified as having BAT. These patients were all male and did not undergo warming (group A). As pediatric patients are known for their predisposition to BAT, a subanalysis was performed on the adult population, with no alteration to the outcomes outlined above being noted.

The distribution of BAT among patients predominantly demonstrated cervical uptake (89%). The supraclavicular area was seen in 75% of patients, the paravertebral in 50%, the suprarenal in 21%, and the paraaortic in 7%. The SUV data were limited by an inadequate population size in the warmed cohort (n = 2) who presented with BAT appearance. Nonetheless, no statistically significant differences were noted between the BAT SUV–to–liver SUV with respect to sex (P = 0.9157), patient age (P = 0.5059), height (P = 0.2940), weight (P = 0.3314), or BMI (P = 0.6756). There was no statistically significant difference in the liver SUV based on either daily maximum temperature (P = 0.5152) or daily minimum temperature (P = 0.4375) supporting its use as a reference.

DISCUSSION

BAT activation is a sympathetic response of the nervous system as a consequence of cold exposure (2). Its appearance on ¹⁸F-FDG PET has been linked to lower temperatures on the days and weeks preceding the PET scan and has been described to be more common in the months succeeding the onset of winter (19). This evidence supports the theory that BAT activation occurs in response to prolonged cold exposure. Nonetheless, BAT has also been visualized on ¹⁸F-FDG PET after cold exposure over a shorter time frame, ranging from 1 or 2 d to several hours before a PET scan (8,15). Whether the activation of BAT occurs as a consequence of acute cold exposure, prolonged cold exposure, or a combination of both remains debatable. Nevertheless, for the potential reduction of BAT on ¹⁸F-FDG PET through the use of warming techniques, it is the acute effects of cold exposure that are the most manageable.

Several methods have been evaluated as a means to prevent the acute effects of cold exposure causing BAT appearance on ¹⁸F-FDG PET. Warming techniques have been included in the patient instructions, such as keeping warm for 48 h before the scan using winter-type clothing, warming the interior of the car before travel to the PET center, and using prewarmed blankets and temperature-controlled rooms for up to 2 h before ¹⁸F-FDG injection (16,17). The warming techniques described have demonstrated effectiveness in reducing the appearance of BAT in controlled research environments. In the clinical setting, it may be difficult to control the patients’ environmental temperature for an extended time. Universal compliance is more likely to be achieved through shortened, simplified preparation instructions. That is, for thermal control to offer a marginal benefit over other reduction methods, such as pharmaceutical approaches, the technique needs to be readily applied in a universal manner with minimal intrusion or demands that might undermine compliance.

The approach adopted in this investigation relied on the use of 2 minimally intrusive techniques: instructing the patient to dress warmly (“in winter-type clothing”) before presenting to the department on the day of the PET scan, and providing warmed blankets and an emergency “space” blanket for insulation during the postinjection uptake phase. These approaches caused minimal interruption to department workflow. Adopting this approach, a 78% reduction in the incidence of BAT on PET scans was observed (15.3% incidence reduced to 3.3%) (Fig. 2). By sex, a 100% reduction in BAT presentation was observed for female patients and 67% for male patients.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

A 28-y-old man with desmoplastic small-round-cell tumor presented in October 2009 for initial ¹⁸F-FDG PET study (left) and subsequently in January 2010 (right). Both studies used approximately 270 MBq of ¹⁸F-FDG and were obtained after 60-min uptake phase. Initial image was obtained without warming techniques. Follow-up image was obtained with warming intervention. Apart from disease progression, presence and absence of significant BAT accumulation is noted.

The predominance of BAT in women is widely reported, and this is evident in the results of previous investigations into the use of thermal control (16,18). Within this study, a higher number of male patients was observed to have BAT. The difference was not deemed significant and was attributed to a male dominance in the random selection process. We also noted that female patients were more likely to respond to thermal control. It is possible that due to the higher percentage of BAT in women than in men, in women the response to intervention occurs over a shorter time.

Study Limitations

Because of the contiguous recruitment of the study and the clinical nature of the PET department, a degree of variability in the 60-min uptake period after radiopharmaceutical administration was encountered. Consequently, the SUV data, despite being concordant with observer data, may lack precision. Furthermore, there are known inaccuracies when SUV is calculated from time-of-flight reconstructions. These limitations were addressed using the BAT-to-liver ratios.

The cohorts were deliberately sampled during seasonally variant periods. Group A (not warmed) was studied during the cooler period of winter and early spring, whereas group B (warmed) was studied during the warmer summer period. From the data collected, there is no way to eliminate seasonal change as the only reason for the change in incidence of BAT. Indeed, it is possible, although unlikely, that the presence of BAT simply represents the chronic thermal environment previously discussed. The specific aim of this investigation was to determine whether the acute thermal environment could be managed. The lack of correlation between incidence of BAT and SUV ratio with daily maximum temperatures in the presence of a statistically significant correlation with the daily minimum temperature suggests that acute thermal regulation played a role in the observed results. Furthermore, in colder periods (group A), patients are more inclined to dress warmly; increase home, office, and car environmental temperatures; and present for the procedure in a warm state. This again supports the role of in-department thermal control in minimizing BAT. The cited temperatures themselves are unlikely to represent the actual temperature of the environment that the patient predominantly occupied during the chronic or acute phases of the study.

Recommendations

Further evaluation is required in a larger population with matching of seasonal variation as a variable between the 2 cohorts. Recording and correlation of patient medications (e.g., β-blockers) should also be included in future studies. Although lacking clinical practicality, there would be value in undertaking a repeated-measures design for more rigorous control.

• Received for publication September 28, 2011.
• Accepted for publication December 12, 2011.